Alteration of serum amino acid profiles by dietary adenine supplementation inhibits fatty liver development in rats

Non-alcoholic fatty liver disease (NAFLD) is a spectrum of disorders progressing from simple steatosis (NAFL) to non-alcoholic steatohepatitis (NASH), cirrhosis, and hepatocellular carcinoma, and is linked to imbalanced macronutrient intake. Beyond lipids and sugars, protein and amino acid nutrition, as well as nucleobases, strongly influence hepatic lipid accumulation. Low-protein or low-arginine diets induce fatty liver in rodents, but prior work indicates dietary amino acid intake per se does not directly correlate with hepatic TAG; rather, comprehensive serum amino acid profiles correlate with liver TAG. Because prior studies largely established correlations, the causal role of serum amino acid profiles remained unresolved. This study tests the hypothesis that serum amino acid composition causally regulates hepatic TAG accumulation and that dietary adenine, known to prevent low-arginine diet-induced steatosis, acts by altering serum amino acid profiles rather than via direct purine metabolic effects.

Previous reports link dietary imbalance of proteins/amino acids and nucleobases to hepatic lipid accumulation in rodents and humans. Low-protein or low-arginine diets induce steatosis, and orotic acid promotes fatty liver, while adenine supplementation blocks low-arginine-induced steatosis. However, mechanisms are unclear, and increased hepatic orotic acid under arginine deficiency was proposed but not definitively established as causative. Serum amino acid profiling has diagnostic value in obesity, diabetes, cancers, and NAFLD; earlier work by the authors showed that comprehensive serum amino acid profiles, rather than single amino acids, correlate well with hepatic TAG across varied amino acid diets in rats. Nonetheless, these were correlation analyses, leaving causality open. Reports on specific amino acids (arginine, methionine, threonine, tryptophan, glutamine, BCAAs) show context-dependent and sometimes contradictory effects on NAFLD, suggesting complexity beyond individual amino acid intake.

Design and diets: Male Wistar rats (5–6 weeks old) were pre-fed a control amino acid-mixed diet (CN; 15% w/w amino acids) and then assigned for 7 days to CN or low-arginine (ΔArg) diets, with or without supplementation of purines/metabolites (adenine, guanine, adenosine, AMP, or IMP) at 3 g/kg unless noted. Additional interventions included diets with combined deficiencies: ΔArg/ΔMet (methionine reduced to one-third), ΔArg/ΔBCAA (valine, leucine, isoleucine each one-third), or ΔArg/ΔHis (histidine one-third). A high IMP supplementation diet (9 g/kg) and intravenous IMP injections (80 or 160 mg/kg on days 1, 3, 5) were also tested. Animal procedures: Rats were housed individually under controlled temperature, humidity, and 12 h light/dark cycles with ad libitum access to food and water. Body weight and intake were monitored daily. After the feeding period, blood from the carotid artery and liver samples were collected under isoflurane anesthesia and stored at −80 °C. Cell culture: H4IIE-C3 rat hepatoma cells were cultured in DMEM with 10% FBS. For experiments, cells were switched to serum-free media for 24 h. A custom medium mimicking the serum amino acid composition of ΔArg-fed rats was formulated: four amino acids elevated in ΔArg serum (methionine, histidine, glutamine, glutamic acid) were doubled versus control medium, and seven amino acids decreased in ΔArg serum (cysteine/cystine, tryptophan, phenylalanine, tyrosine, arginine, threonine, glycine) were omitted. Measurements: Hepatic and cellular TAG were quantified by enzymatic assay following lipid extraction (modified Folch method). Serum amino acids were measured by LC-MS/MS. Hepatic purine metabolites (adenine, adenosine, AMP, IMP, guanine, guanosine, hypoxanthine, xanthine) were quantified by LC-MS/MS. TAG secretion assay: After 7-day diets, fasting rats received tyloxapol intravenously; serial serum TAG was measured at 0–4 h. The slope of TAG increase (under lipoprotein lipase inhibition) indexed hepatic TAG secretion rate. Machine learning (MLP): A multilayer perceptron trained previously on 95 rats (21 inputs: 20 amino acids + cystine; output: liver TAG) was used without modification (Keras 2.3.1/TensorFlow 2.1.0). Architecture: 4 hidden layers (300/300/300/100 neurons) with ReLU activations; dropout 0.5/0.4/0.4/0.4/0.3; early stopping up to 10,000 epochs; loss MSE, optimizer RMSprop. The trained model predicted liver TAG from new serum amino acid profiles (purine-supplemented groups). Statistics: Data are mean ± SEM. Two-group comparisons used Student’s t-test; multiple groups used one-way ANOVA with Tukey–Kramer post hoc when ANOVA p<0.05; significance at p<0.05.

• Low-arginine (ΔArg) diet rapidly induced hepatic TAG accumulation in rats within 7 days and altered serum amino acid profiles: increased methionine, histidine, glutamine, glutamic acid; decreased cysteine/cystine, tryptophan, phenylalanine, tyrosine, arginine, threonine, glycine.
• A serum-free culture medium mimicking the ΔArg serum amino acid profile significantly increased TAG in H4IIE hepatoma cells versus control medium (p=0.000115), indicating that the specific extracellular amino acid profile is sufficient to drive hepatocellular TAG accumulation in a cell-autonomous manner.
• Among purines and related metabolites tested (adenine, guanine, adenosine, AMP, IMP), only dietary adenine supplementation (3 g/kg) abolished ΔArg-induced hepatic TAG accumulation (p<0.0001 vs ΔArg) and fully rescued hepatic TAG secretion in vivo (tyloxapol assay), whereas guanine did not.
• Hepatic purine metabolite analysis showed that hepatic adenine levels did not correlate with liver TAG; adenine supplementation elevated hepatic IMP, which negatively correlated with liver TAG, but heavy dietary IMP (9 g/kg) only slightly reduced TAG and intravenous IMP (80–160 mg/kg) had no effect, arguing against purine metabolite levels as the primary mechanism of protection.
• A trained multilayer perceptron using only serum amino acid concentrations accurately predicted the suppression of hepatic TAG by adenine supplementation, implying that adenine’s effect is mediated through changes in serum amino acid profiles.
• Serum amino acid profiling revealed that ΔArg, ΔArg+adenosine, ΔArg+AMP, and ΔArg+IMP groups shared similar profiles associated with fatty liver, whereas ΔArg+adenine exhibited a distinct profile aligned with non-steatotic outcomes.
• Guided by these differences, reducing methionine or BCAAs (valine, leucine, isoleucine) to one-third in the ΔArg diet markedly altered serum amino acid profiles and significantly prevented hepatic TAG accumulation; histidine reduction had little effect.

The study demonstrates a causal role for comprehensive serum amino acid profiles in regulating hepatic lipid metabolism. A ΔArg diet creates a distinct serum amino acid signature that directly induces hepatocellular TAG accumulation, as recapitulated in serum-free hepatocyte culture. Dietary adenine uniquely prevents ΔArg-induced steatosis not by increasing hepatic adenine or downstream purines per se, but by reshaping the serum amino acid profile into a non-inducible pattern for lipid accumulation. Machine learning predictions based solely on serum amino acids accurately identified adenine’s protective effect, reinforcing the primacy of the amino acid milieu over individual compounds in purine metabolism. Intervention diets showed that lowering methionine or BCAA availability within a ΔArg background abolishes hepatic TAG accumulation, highlighting the importance of relative amino acid concentrations and combinations rather than absolute intakes of single amino acids. Canonical lipid regulatory pathways (PPARα, SREBP1c) were not affected by amino acid deficiency in prior observations cited by the authors, suggesting alternative sensing and signaling mechanisms through which hepatocytes transduce extracellular amino acid profiles into lipid metabolic outcomes. These insights link nutrient-derived amino acid signals to NAFLD pathophysiology and support the development of serum amino acid profile-based diagnostics and dietary interventions.

Comprehensive serum amino acid profiles are not merely correlates but causative regulators of hepatic TAG accumulation. Low-arginine diets induce a serum amino acid signature that suppresses hepatic TAG secretion and promotes steatosis; dietary adenine prevents fatty liver by altering this signature. Modulating specific components of the profile—particularly methionine and BCAAs—can abolish ΔArg-induced hepatic TAG accumulation in vivo. These findings introduce the concept of a metabolic regulatory amino acid signal governing hepatic lipid metabolism and suggest practical avenues for prediction and prevention of NAFLD using serum amino acid profiling and targeted dietary strategies. Future research should delineate the specific amino acid combinations and relative ratios that drive hepatic lipid regulation, identify the hepatocellular sensors and signaling pathways involved, and evaluate translatability to human NAFLD.

Mechanistic details remain unresolved: which precise amino acid combinations and relative concentrations are necessary and how hepatocytes sense and transduce these extracellular amino acid signals are unknown. The protective effect of adenine is not explained by measured hepatic purine metabolites, leaving the intermediary mechanisms unclear. Findings are based on short-term (7-day) interventions in male Wistar rats; generalizability to other models and to humans was not assessed. Not all amino acids appear equally important (e.g., histidine reduction had little effect), but the minimal feature set was not defined. Certain NAFLD etiologies (e.g., hepatitis C virus-associated) may not be directly linked to serum amino acids, limiting broad applicability.